Swordfish vs Devastator

Ad: This forum contains affiliate links to products on Amazon and eBay. More information in Terms and rules

KNOTS!!...maybe that's what it was, and not mph.
90 kts is almost 104 mph, so that's a little faster.
From what I'm learning here, it seems the torpedoes had wood shrouds over the nose and tail that helped protect and stabilize the torpedo in the air. They sheared off upon impact, but they also slowed the torpedo slightly and that helped them home in on the target.
I guess it was this that allowed launch speeds to increase.
Not only launch speed but height too, the ability of the torpedo to keep straight and get itself to the correct depth had to be improved too.
 
Hey SaparotRob,

If you can find it (and have not already seen it), there is a very good article on the serious problems with the US torpedoes in the early- to mid-war. It involved all of the new generation of torpedoes (ie the Mk 13 aircraft, Mk14 submarine, and Mk 15 surface ship). When testing was finally done in late-1943(?) it was found that there was a 130% failure rate of the Mk 13. This was due to multiple failure modes for multiple torpedoes. IIRC less than 20% of the torpedoes tested hit their target and exploded. I had a PDF copy of the article on my last computer that crapped out on me about a year ago, otherwise i would post it (it was originally published for government use so the majority of the article is public domain). If you want to search for it, I first ran across it in the Naval Institute Press magazine 'Proceedings' and it had the words "Torpedo Scandal" in the title of the article. There is another article online with the words"Great Torpedo Scandal" which covers some of the same info but is not (I think) the one in Proceedings.

The Mk 13 Mod 0 (the last of which were used up in the early-war battles) had fewer problems (during test pre-war) than the later mods, and the Mod 1 had fewer problems than the Mod 2, etc. . . until whatever Mod was the successful late-war variant.
 
Last edited:
Can someone tell me how you get a 130% failure rate?
I think it was in the post, when there are many ways things can go wrong a single torpedo can fail in three or four ways at the same time. Statisticians have mysterious ways to perform their wonders. As a logical expression it is nonsense, but as a statistical expression it points to many big problems
 
They would probably get 130% failure rate by examining component failure modes, then adding them together. Failure analysis was, rather obviously, in its childhood as this is invalid reasoning.

Actually analyzing failure statistics can result in some interesting results. Insurance companies do it all the time, which is why, for example, the insurance rates (as a percentage of aircraft cost) are higher for genav pilots flying twins. While one would think that a twin would crash less, and this is certainly true for highly experienced transport pilots, it's less true for many private pilots who both have less overall experience, fly less often, and have less access to advanced training aids. Another counter-intuitive result is that twin-turbine helicopters (at least in the 1980s, the last time I was actively keeping track) autorotate more often than single turbine helicopters, as the rate of gearbox failure in twins was greater than the combined rate of engine plus gearbox failure in single turbine helicopters. I wonder how the loss rate for the F-18 compares with that of the F-16, as I keep seeing news about F-18s crashing after losing one engine.
 
They would probably get 130% failure rate by examining component failure modes, then adding them together. Failure analysis was, rather obviously, in its childhood as this is invalid reasoning.

.
Well that depends on your reasoning. There are many ways to present the results of an investigation. Saying there is a 130% failure rate on one of the two main weapons that the aircraft carrier fleet was built for will grab peoples attention. It says that a ship and crew and its pilots are taking huge risks at massive cost to fire something that statistically has no chance of working, the few that do hit the target and explode as planned are lucky quirks.
 
Hey SparrotRob,

If you can find it (and have not already seen it), there is a very good article on the serious problems with the US torpedoes in the early- to mid-war. It involved all of the new generation of torpedoes (ie the Mk 13 aircraft, Mk14 submarine, and Mk 15 surface ship). When testing was finally done in late-1943(?) it was found that there was a 130% failure rate of the Mk 13. This was due to multiple failure modes for multiple torpedoes. IIRC less than 20% of the torpedoes tested hit their target and exploded. I had a PDF copy of the article on my last computer that crapped out on me about a year ago, otherwise i would post it (it was originally published for government use so the majority of the article is public domain). If you want to search for it, I first ran across it in the Naval Institute Press magazine 'Proceedings' and it had the words "Torpedo Scandal" in the title of the article. There is another article online with the words"Great Torpedo Scandal" which covers some of the same info but is not (I think) the one in Proceedings.

The Mk 13 Mod 0 (the last of which were used up in the early-war battles) had fewer problems (during test pre-war) than the later mods, and the Mod 1 had fewer problems than the Mod 2, etc. . . until whatever Mod was the successful late-war variant.
Thanks for the heads-up. I'll check that out.
Have you seen "Failure is Like Onions" on Drachinifel's YouTube site? An entertaining history of the whole MK 14 scandal.
 
I think it was in the post, when there are many ways things can go wrong a single torpedo can fail in three or four ways at the same time. Statisticians have mysterious ways to perform their wonders. As a logical expression it is nonsense, but as a statistical expression it points to many big problems
Thanks for this, I always did tend to think of things simplistically.
 
Can someone tell me how you get a 130% failure rate?

This is from the official history of the USN Bureau of Ordnance:


"Despite the complications that were attending the other phases
of torpedo development, the Bureau of Ordnance considered the
aircraft torpedo problem so important that it was assigned the
highest priority at the Newport Station. The improvements and
modifications of 1942 and 1943 still left the weapon unpopular,
however, and production problems were as great as those stemming
from incomplete development. In mid-1943 an analysis of

105 torpedoes dropped at speeds in excess of 150 knots showed
clearly why aviators distrusted the Mark 13: 36 percent ran cold,
20 percent sank, 20 percent had poor deflection performance, 18
percent gave unsatisfactory depth performance, 2 percent ran on
the surface, and only 31 percent gave a satisfactory run. The total
in excess of 100 percent proved that many torpedoes were subject

to more than one of the defects, just as the bulk of the problems
were still due to the effects of poor air stabilization on water behavior.
Better performance at reduced aircraft speeds was small
comfort since aviators could not be held down by paper restrictions
that imposed serious and dangerous handicaps in combat. And
even when they accepted the limitations, the water entry behavior
of the torpedo produced frequent hooking and broaching.
Time promised to complicate the problem still further. Unsatisfactory
for existing planes, the torpedo would certainly fail to
utilize the potentialities of aircraft then under development."
 
Thanks for this, I always did tend to think of things simplistically.
Number crunching throws up these things. using imaginary numbers, if you test 1000 guns and two fail due to jamming, one fails because of the feed, one because it didn't reload and one just didn't fire, then it wouldn't be seen as unreasonable to add them together and say 5/1000 or 0.5% were failures. Only when the numbers increase does it become a sort of nonsense. The quoted rate of 130% should be compared to RCAFsons post above where the actual rate in practice was 31% being successful. If you solve one mode of failure and run the test again you may well have more success but you could have less, with a different distribution of reasons to fail.
 
Let's say a torpedo's guidance fouls up and its detonator fails, too, would that torpedo have a 200% failure rate?
 
Let's say a torpedo's guidance fouls up and its detonator fails, too, would that torpedo have a 200% failure rate?
That is exactly what the statistics mean, not 200% but two failures in the 100 torpedoes tested. A similar and more common expression concerns bomber crews and I have just read it in a history of 617 squadron. If bomber aircraft losses were on average 5% per mission then on a 30 mission tour a crew has a 150% chance on being lost. However 617 squadron was composed almost entirely of crews that finished a 30 mission tour. This shows that losses were not random, the best pilots, the best navigators engineers and gunners improved their chances of survival in many ways. Those lost were frequently on their first or early missions, and their replacements were lost in similar fashion, at the other end of the scale Guy Gibson had completed 3 operational tours before he joined 617 squadron.
 
Had been the other way round, and all the Zeros were at HA beating the hell out of the SBDs and thus not noticing the approaching TBDs at LA, the Devastators may have scored crippling hits. The AA on IJN was apparently rubbish.


I'd also argue the TBD had success at Midway. Were it not for the Devastators drawing the Zeros down to LA it's not assured that the Dauntless strike from HA would have got into position unmolested over the KB to strike the killer blows. It's a team effort, each USN aircraft played a part.

Actually no, in Shattered Sword Parshall and Tully bust that myth by actually working out chronologically all the attacks on Kido Butai. VT-8 was shredded an hour before the decisive SBD attack, and point out that their contribution was the same as VT-6 - "disrupting the counteroffensive activities of the Japanese carriers".

Although I totally agree with the bolded parts of your posts, it was a team effort, let's not forget the USAAF sent B-17's and B-26's out, which, while they may not have scored any hits, they also disrupted carrier operations prior to the killer strike. Including coming within a whisker of making Nagumo a B-26 hood ornament, and while I'd like to take credit for that witty turn of phrase, it's straight out of Shattered Sword.

On the Swordfish, I agree, I've referenced that link BPF more than once here, and I think the Swordfish would have given a good account of itself in a predawn or night time attack. You're right, Kido Butai for all its supposed invincibility, would have been hard pressed against that kind of attack.
 
VT-8 was shredded an hour before the decisive SBD attack, and point out that their contribution was the same as VT-6 - "disrupting the counteroffensive activities of the Japanese carriers".

Although I totally agree with the bolded parts of your posts, it was a team effort, let's not forget the USAAF sent B-17's and B-26's out, which, while they may not have scored any hits, they also disrupted carrier operations prior to the killer strike.
If there's one thing the Kido Butai should have done is to operate all four carriers as an organic unit, by which I mean stop operating each carrier as a stand-alone combatant. Instead of each fighter having to return to its home carrier to replenish ammunition and fuel and thus block counterstrike launches, assign one carrier for fighter rotation and the other three carriers for strike launch. If the fighter carrier is disabled, rotate the fighter rotation to another carrier, whilst the two remaining continue strike prep.
 
If there's one thing the Kido Butai should have done is to operate all four carriers as an organic unit, by which I mean stop operating each carrier as a stand-alone combatant. Instead of each fighter having to return to its home carrier to replenish ammunition and fuel and thus block counterstrike launches, assign one carrier for fighter rotation and the other three carriers for strike launch. If the fighter carrier is disabled, rotate the fighter rotation to another carrier, whilst the two remaining continue strike prep.
You would loose some efficiency though, no? I mean, four carriers can assemble a strike force faster than three. And they also can bring up a bigger fighter screen in the same time.
 
If the IJN had allowed composite air groups instead if dedicating an air group to each carrier, then they could have combined the air groups of Shokaku and Zuikaku to bring Zuikaku up to strength and had her back in action with the rest of Kido Butai.
As it stands, they should have never allowed Carrier Division 1 and 2 to become seperated from the main task Force as they did.
The additional scouting aircraft of the 3 battleships and several cruisers would have proven invaluable along with the concentrated AA - plus the two light carriers Zuiho and Hosho would have provided 12 A6M2, 12 B5N2 and 8 B4Y1 aircraft that could have certainly contributed.
 

Users who are viewing this thread

Back